{"id":3825,"date":"2016-03-29T11:30:04","date_gmt":"2016-03-29T16:30:04","guid":{"rendered":"https:\/\/skepticink.com\/backgroundprobability\/?p=3825"},"modified":"2016-03-29T11:19:53","modified_gmt":"2016-03-29T16:19:53","slug":"graph-week-modeling-democratic-primaries","status":"publish","type":"post","link":"https:\/\/skepticink.com\/backgroundprobability\/2016\/03\/29\/graph-week-modeling-democratic-primaries\/","title":{"rendered":"Graph of the Week \u2014 Modeling the Democratic Primaries"},"content":{"rendered":"<p>A couple of days ago,\u00a0a friend tipped me off to the following scatter plot by Will Jordan, illustrating the race between Clinton and Sanders:<\/p>\n<blockquote class=\"twitter-tweet\" data-width=\"500\" data-dnt=\"true\">\n<p lang=\"en\" dir=\"ltr\">Bernie Sanders&#39;s performance broken down by statewide %white population and whether contest was a caucus\/primary <a href=\"https:\/\/t.co\/uVStpEyyqf\">pic.twitter.com\/uVStpEyyqf<\/a><\/p>\n<p>&mdash; Will Jordan (@williamjordann) <a href=\"https:\/\/twitter.com\/williamjordann\/status\/714154078893391872?ref_src=twsrc%5Etfw\">March 27, 2016<\/a><\/p><\/blockquote>\n<p><script async src=\"https:\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script><\/p>\n<p>This was followed by some good discussion about regression and r-squared measures for goodness of fit.<\/p>\n<p>It occurred to me that the percentage of &#8220;Non-Hispanic White&#8221; (Lord, how I hate that phrase) is basically acting as negative correlate for the Black and Latino vote, which is why Alaska and Hawaii seem like outliers in his plot. Accordingly, I recreated the graph using combined\u00a0percentages of <a href=\"http:\/\/www.census.gov\/prod\/cen2010\/briefs\/c2010br-06.pdf\" target=\"_blank\">African Americans<\/a> and <a href=\"http:\/\/www.census.gov\/prod\/cen2010\/briefs\/c2010br-04.pdf\" target=\"_blank\">Latinos<\/a> by state.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-3826\" src=\"https:\/\/skepticink.com\/backgroundprobability\/files\/2016\/03\/Clinton_Vote_Demographics.png\" alt=\"Clinton_Vote_Demographics\" width=\"635\" height=\"556\" srcset=\"https:\/\/skepticink.com\/backgroundprobability\/files\/2016\/03\/Clinton_Vote_Demographics.png 635w, https:\/\/skepticink.com\/backgroundprobability\/files\/2016\/03\/Clinton_Vote_Demographics-300x263.png 300w\" sizes=\"auto, (max-width: 635px) 100vw, 635px\" \/>Even after seeing Jordan&#8217;s scatterplot, I was surprised at how tight the pattern is here. Nearly 60% of the variance in Clinton\u00a0support can be explained by simple linear model based on the fraction of the population which falls into just two categories.\u00a0She has yet to lose the popular vote when\u00a0those categories exceed 25% combined.<\/p>\n<p>If this model holds up\u2014of which I make no assurances\u2014what would it tell us about the delegate-rich races coming up in New York (247 pledged delegates), Maryland (95\u00a0pledged delegates), Pennsylvania (189\u00a0pledged delegates), California (475\u00a0pledged delegates), and New Jersey (126\u00a0pledged delegates)? New York is 15.9% Black, 17.6% Latino; Maryland is 29.4% Black, 8.2% Latino; Pennsylvania is 10.8% Black, 5.7% Latino; California is 6.2% Black, 37.6% Latino; and New Jersey is 13.7% Black, 17.7% Latino, according to the same census data upon which the above plot was built. If the model does prove predictive, Clinton would likely win all of these large states except for Pennsylvania, which may well go to Sanders.<\/p>\n<p>All the usual caveats apply; no doubt there are vastly more refined ways to model the nominating contests based on underlying demography. If you happen to find one in the wild, please do let me know.<\/p>\n<p>&nbsp;<\/p>\n<p>&nbsp;<\/p>\n<p>&nbsp;<\/p>\n<p>&nbsp;<\/p>\n","protected":false},"excerpt":{"rendered":"<p>A couple of days ago,\u00a0a friend tipped me off to the following scatter plot by Will Jordan, illustrating the race&hellip;<\/p>\n","protected":false},"author":23,"featured_media":3826,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[18,10],"tags":[],"class_list":["post-3825","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-damned-lies-and-statistics","category-politics"],"_links":{"self":[{"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/posts\/3825","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/users\/23"}],"replies":[{"embeddable":true,"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/comments?post=3825"}],"version-history":[{"count":0,"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/posts\/3825\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/media\/3826"}],"wp:attachment":[{"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/media?parent=3825"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/categories?post=3825"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/skepticink.com\/backgroundprobability\/wp-json\/wp\/v2\/tags?post=3825"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}